Table 1.
Descriptive analysis of the dataset variables.
Fig 1.
Distribution of tropes in films (log-log scale applied, zeros were treated adding one to all the values).
Almost all the films have less than 200 tropes. The higher number of tropes in a film is 677.
Fig 2.
Rating distribution of all films in the dataset.
Skewness: -1.51. Kurtosis: 3.7.
Fig 3.
Distribution of votes (popularity) in films (log-log scale is applied, zeros were treated adding one to all the values).
Table 2.
Descriptive analysis of votes and ratings by periods for all films of the dataset.
Notice that the data for the period (2000, 2020] is not complete.
Fig 4.
Number of films by year in the dataset extracted from TV Tropes.
The dataset does not include all films of 2019, as it was generated before the year ended.
Table 3.
Percentage of film genres in the dataset.
NOTE: A film can have more than one genre.
Fig 5.
Jaccard index heatmaps of the overlapping of tropes.
“\N” is used to denote that a film does not belong to a genre. a) Heatmap of the overlapping of tropes with respect to the top 100 rated films. b) Heatmap of the overlapping of tropes with respect to the top 100 voted films. c) Heatmap of the overlapping of tropes with respect to all the tropes of the genres.
Table 4.
Performance measures from 10 arbitrary communities obtained with the Leiden algorithm for the co-tropes network.
Fig 6.
Strategic diagram of the communities obtained from the co-tropes network.
Depending on centrality (x-axis) and density (y-axis) of each community, they can be placed in any of the 4 quadrants whose center is in (0.5,0.5). From left to right and top to bottom: specialized/peripheral, motor (central/developed), emerging/decaying and transversal. Size of the dots is proportional the number of movies in each community (also shown in numeric value).
Fig 7.
Distribution by decade of the movies of all communities detected in the emerging/dissapearing quadrant of the strategic diagram.
Fig 8.
Co-tropes spanning tree of the co-occurrence network.
The larger the node more films have the trope. The larger the edges, more films both tropes have in common.
Table 5.
Rate of genres for 10 arbitrary communities obtained with the algorithm.
Table 6.
Co-films performance measures.
Table 7.
Co-films rate of genres.
Fig 9.
Co-films spanning tree of the co-occurrence network.
The larger the node more tropes appear in the film. The larger the edges, more tropes both films have in common.